British Journal of Political Science, Forthcoming. James L. Gibson Sidney W. Souers Professor of Government

Transcription

1 THE SUPREME COURT AND THE US PRESIDENTIAL ELECTION OF 2000: WOUNDS, SELF-INFLICTED OR OTHERWISE? British Journal of Political Science, Forthcoming James L. Gibson Sidney W. Souers Professor of Government Department of Political Science Washington University, St. Louis Campus Box 1063, 219 Eliot Hall St. Louis, MO United States Gregory A. Caldeira Distinguished University Professor University Chaired Professor of Political Science Department of Political Science Ohio State University 2140 Derby Hall, 154 N. Oval Mall Columbus, Ohio United States Lester Kenyatta Spence Assistant Professor Department of Political Science Washington University, St. Louis Campus Box 1063, 219 Eliot Hall St. Louis, MO United States Version 3.3, 18 October 2002 *This project would not have been possible without the support of the Weidenbaum Center on the Economy, Government, and Public Policy, Washington University, St. Louis, and The Ford Foundation (Grant Number ). We are especially indebted to Steve Smith, Director of the Weidenbaum Center, for his encouragement of this work. An earlier version of this paper was presented at the American Politics Workshop series at Columbia University. Professors Roger E. Hartley, Robert Erikson, and David Epstein made most useful comments on an earlier version of this paper. We also appreciate the research assistance of Eric Lomazoff and Hannah Pierce of the Russell Sage Foundation.

2 Abstract The conventional wisdom about the US Supreme Court and the 2000 presidential election is that the Court wounded itself by participating in such a partisan dispute. By wounded people mean that the institution lost some of its legitimacy. Evidence from our survey, conducted in early 2001, suggests little if any diminution of the Court s legitimacy in the aftermath of Bush v. Gore, even among African Americans. We observe a relationship between evaluations of the opinion and institutional legitimacy, but the bulk of the causality seems to flow from loyalty to evaluations of the case, not vice versa. We argue that legitimacy frames perceptions of the Court opinion. Furthermore, increased awareness of the activities of the Court tends to reinforce legitimacy by exposing people to the powerful symbols of law. In 2000, legitimacy did indeed seem to provide a reservoir of good will that allowed the Court to weather the storm created by its involvement in Florida s presidential election.

3 The US presidential election of 2000 reminds us once again of the importance of the legitimacy of political institutions. Consider this highly simplified view of the election. A bitter political controversy arises. The dispute bounces around various institutions, with no definitive resolution. Finally, the US Supreme Court intervenes and makes a decision. Some grumble about the ruling, but political elites call for it to be respected, much if not most of the mass public seems to accept the decision, and the brouhaha ends. The country gets on with its business. Scholars then talk of the Court expending its political capital, and wonder about the efficacy of the Court in future clashes, but the Court seems to have enough legitimacy to get away with its ruling and make it stick. The circumstances surrounding Bush v. Gore 1 may well enter our textbooks one day as a stellar example of the power and efficacy of institutional legitimacy. The Supreme Court decision settling the outcome of the US presidential election unleashed a torrent of criticism. Not only was the logic of the opinion assailed 2 but many also judged the Court s opinion, and even the Court itself, as illegitimate. 3 For instance, 585 law professors placed an advertisement in the New York Times on 13 January 2001, condemning the Court s decision as illegitimate. 4 Perhaps for the first time, public opinion pollsters queried the American people using words like legitimate and legitimacy. 5 Scholars complained about a self-inflicted wound on the Court, 6 as many questioned whether Bush v. Gore undermined the ability of the Court to rule on controversial issues in the future. In short, in the eyes of many, Bush v. Gore subtracted mightily from the institutional legitimacy enjoyed by the US Supreme Court. A great deal of hyperbole surrounded the US election, and politicians, professors, and pundits made many outlandish empirical claims at the time. Terms like legitimacy were nearly -1-

4 always used loosely, without rigorous definition. Unfortunately, because serious scholarly inquiry into the Court s legitimacy has been limited, we have little systematic knowledge about how the election may actually have influenced the legitimacy of the Court. The theoretical issues raised by this dispute are not, however, unfamiliar to scholars. Researchers have long been interested in the question of how an institution s performance affects its legitimacy, and in particular how the Supreme Court s decisions shape people s views of it. 7 Still, the relationship between judgments of particular decisions and more general attitudes toward political institutions like the Supreme Court is under-analysed and poorly understood within political science. How legitimate is the US Supreme Court today, and how did the 2000 election affect the legitimacy of the institution? Though the question of change in support is difficult to answer, we adduce some new evidence on the Court s legitimacy in the period shortly after the conclusion of the presidential election. One purpose of this article is thus to provide empirical results on how Americans viewed the Court in the aftermath of the 2000 presidential election, based on a nationally representative sample interviewed at the beginning of We have several additional objectives as well. First, because we draw our indicators of legitimacy from extant research, we are able to assess the degree to which the Court s legitimacy changed following its decision in Bush v. Gore. We do not have panel data, but we do have cross-sectional data from three surveys. Second, because our survey includes an over-sample of black Americans, we are able to examine the attitudes African Americans hold toward the Court, both now and in comparison to earlier surveys. Third, we employ a variety of statistical techniques to untangle the causal relationship between views of the opinion in Bush v. Gore and more general loyalty to the Court as an institution. Finally, since our results are quite contrary to conventional wisdom, we reconsider extant theories of how people update their views toward -2-

5 political institutions, focussing in particular on the framing associated with incidents in which courts become salient. We begin with a brief review of Legitimacy Theory. THE THEORY OF INSTITUTIONAL LEGITIMACY Considerable agreement exists among social scientists on most of the major elements of Legitimacy Theory. For instance, most agree that legitimacy is a normative concept, having something to do with the right (moral and legal) to make decisions. Authority is often used as a synonym for legitimacy. Legitimate institutions are those with an authoritative mandate to render judgments for a polity. Under the influence of Easton, 8 scholars use diffuse support as a synonym for legitimacy. The concept refers to a reservoir of favourable attitudes or good will that helps members to accept or tolerate outputs to which they are opposed or the effects of which they see as damaging to their wants. 9 Diffuse support is institutional loyalty, support not contingent upon satisfaction with the immediate outputs of the institution. Most analysts distinguish between approval of policy outputs in the short-term and more fundamental loyalty to an institution over the long-term. Specific support is satisfaction with the immediate policy outputs. Institutions without a reservoir of good will may be limited in their ability to go against the preferences of the majority, even when it is necessary or wise to do so. 10 Thus, a crucial attribute of political institutions, and courts in particular, is the degree to which they enjoy the loyalty of their constituents. The Dynamics of Opinion: How Do People Update Their Views? One of the most important unanswered questions for the field has to do with whether pre-existing loyalty to the Court cushions the effect of unpopular decisions (such as Bush v. Gore), or whether highly charged decisions undermine an institution s basic legitimacy. Legitimacy -3-

6 Theory obviously suggests the former, but some important recent research suggests the latter. For instance, Grosskopf and Mondak examine whether confidence in the Supreme Court derives solely from stable factors such as core democratic values, or if citizens alter their evaluations to take into account their views of the Court s ruling. 11 They rightly note the importance of the answer to this question: If only core values matter, then a static depiction of support treating it as a virtually inexhaustible resource can be justified. However, if specific decisions can be shown to have an impact on support, then a dynamic view of legitimacy is more appropriate. 12 They conclude that confidence in the Court is very much a function of perceptions and evaluations of court opinions, and that unpopular decisions erode the institution s political capital. This finding directly challenges one of the central propositions of Legitimacy Theory. At the level of the individual citizen, few longitudinal studies have investigated how information about the Supreme Court influences overall evaluations of the institution. Perhaps the best of these studies is Hoekstra s analysis of local reactions to Supreme Court opinions directly affecting several local communities. 13 She discovers that satisfaction with [Court] decisions influences subsequent evaluations of the Court. 14 Unfortunately, however, the conclusions of most earlier studies are limited because they employ indicators of Court support that are not valid measures of institutional legitimacy. For instance, Hoekstra s measure is: In general do you approve or disapprove of the way the Supreme Court is handling its job? Do you approve/disapprove very strongly, strongly, or not strongly? 15 Had this research a more valid measure of legitimacy, it might well have discovered that because the Court has a solid reservoir of good will, unpopular decisions generate ire that dissipates quickly and has no lasting consequence for the legitimacy of the institution, a possibility that Grosskopf and Mondak 16 recognize. -4-

7 Thus, despite the acknowledged theoretical importance of this issue, we have a long way to go in understanding the relationship between institutional performance and legitimacy. Previous research has not been able to overcome measurement limitations, in part because it has not focussed on actual Supreme Court decisions that are both controversial and salient to the American people. 17 The Court s intervention in the Florida election provides an intriguing opportunity to advance our understanding of the connection between policy satisfaction and dissatisfaction and institutional legitimacy. Previous Research on the Impact of Bush v. Gore on Public Evaluations of the Court Most of the polls conducted in the aftermath of the presidential elections used measures of confidence to assess the impact of the election on the Supreme Court. For instance, Harris Interactive used the conventional question employed by the General Social Survey, asking: As far as the people running the US Supreme Court are concerned, would you say you have a great deal of confidence, only some confidence, or hardly any confidence at all in them? The results at the beginning of 2001 reveal only slightly less confidence than the results of a comparable survey conducted at the beginning of Other poll data support a similar conclusion on the question of confidence. Opinions are more evenly split over whether the Court based its decision on law or some combination of partisanship, politics, and the justices own political preferences. A Hart- Teeter/NBC/Wall Street Journal survey, conducted 7-10 December 2000, reported that 34 per cent of the respondents believed that the Court s decision in Bush v. Gore was based mostly on the law but that 53 per cent believe the decision was based mostly on politics. If one were forced to draw a conclusion from these and other fragmentary bits of survey evidence, one might conclude that the Court s ruling in Bush v. Gore diminished confidence in the Court little despite the fairly widespread view that the Court s decision was politically motivated. -5-

8 The most rigorous analysis reported to date is Kritzer s. Based on a tracking poll in the field at the time of the election, he concludes that the Court s involvement in the dispute had measurable consequences for the public s view of the Court, at least in the short term 19 (emphasis in the original). He also argues that views of the Court became more partisan as a result of the election, although his final judgment is that the impact of the ruling is perhaps not as great as many might have expected. Unfortunately, however, Kritzer s conclusions must be treated as not especially relevant to our work since his research was not designed to assess institutional legitimacy and therefore lacks a valid measure of the legitimacy of the Supreme Court. 20 Thus, the limited data available do not produce a clear verdict on whether the Court s intervention in Bush v. Gore had any consequences, disastrous or otherwise, for the legitimacy of the institution. Fortunately, we specifically designed our survey to investigate the question of the linkage between judgments of the Court s decision and the legitimacy of the institution itself (for technical details on the survey see Appendix A). INSTITUTIONAL LOYALTY IN THE AFTERMATH OF THE ELECTION Our thinking about institutional loyalty follows a considerable body of research on conceptualizing and measuring mass perceptions of high courts. 21 That research treats institutional loyalty as opposition to fundamental structural and functional changes in the institution; 22 and it is grounded empirically in the history of attacks by politicians against courts in the US (e.g., court-packing) 23 and elsewhere (e.g., manipulation of appellate jurisdiction). 24 As Caldeira and Gibson describe it, those who have no loyalty toward the Supreme Court are willing to accept, make, or countenance major changes in the fundamental attributes of how the high -6-

9 bench functions or fits into the US constitutional system. 25 Loyalty is also characterized by a generalized trust that an institution will perform acceptably in the future. Loyalty thus ranges from complete unwillingness to support the continued existence of the institution to staunch institutional fealty. Table 1 reports the responses from the six statements we use to measure loyalty toward the US Supreme Court. The first three columns of figures represent the frequencies after collapsing strong and not so strong responses, and the column labelled Supportive of the Court reports the percentage of respondents giving answers indicating loyalty to the institution, regardless of whether loyalty requires an agree or disagree reply. The means and standard deviations are based on the uncollapsed data, and in every instance higher scores indicate more loyalty toward the Supreme Court. [PLACE TABLE 1 ABOUT HERE] These data indicate at least a moderate level of loyalty toward the Supreme Court among most Americans. On average, 3.8 of the statements elicit support for the Court, and 77.7 per cent of the respondents endorse at least three of the six statements (data not shown). Except for the two items on partisanship, strong majorities express support for the Court on four of the six statements. On the clearest measure of institutional loyalty, the first (but also easiest) item, support is extremely widespread: Over four of five Americans assert that it would not be better to do away with the Court, even if there were fairly widespread displeasure with its decisions. Although a significant minority worries about politics and partisanship on the Court, over threefourths of the sample asserts that the Court not the leaders of the Court can generally be trusted. These data indicate that the Supreme Court enjoys a reasonably solid reservoir of good will, even in the aftermath of the tumultuous presidential election of Placing these findings in the context of cross-national research, the US Supreme Court is a fairly legitimate institution

10 The responses to these propositions are positively correlated, with an average inter-item correlation of.26, a decent level of inter-correlation given that these are survey data (with all the general problems of unreliability). Reliability is respectable, with a Cronbach s alpha ( an indicator of internal consistency) of.68. Deletion of none of the variables would increase the alpha coefficient. The statement with the weakest relationship to the total item set (as indicated by its squared multiple correlation coefficient) is the last one, a complicated statement about whether the Court should be subservient to majority opinion. When factor analyzed, these six items generate a single significant factor (the eigenvalue of the second factor extracted is.91), with all of the indicators loading significantly on the factor. We have calculated an Index of Institutional Loyalty as the mean response to these six items. Thus, these measures appear to have excellent psychometric properties. Change in Loyalty Toward the Court It is impossible in a cross-sectional survey to analyse individual-level change in opinions. However, two earlier national surveys employed similar measures of institutional loyalty, so we can derive some indication of the nature of macro-level change. 27 Table 2 reports a comparison of two of the items asked in 2001 with similar questions asked in 1987 in a national survey. 28 Since both studies included representative over-samples of black Americans, and since the election of 2000 had special relevance to black Americans, we report the findings according to the race of the respondent. Although the items put to the respondents are not identical, they are similar enough to be able to draw some inferences about change. 29 [PLACE TABLE 2 ABOUT HERE] Several aspects of this table are important. First, opinions toward the Court in 2001 are considerably more polarized. There are many fewer uncertain or don t know responses to our statements in 2001, surely reflecting the heightened salience of the Court during this period -8-

11 of political controversy. Second, among white Americans, we find no evidence whatsoever of a diminution of support for the Court between 1987 and On the first statement, opinions changed insignificantly, perhaps due to a ceiling effect (opinion in 1987 was already about as positive as it could be). On the second proposition, if anything, support for the Court increased. For instance, 62.8 per cent in 2001 would resist efforts to alter the jurisdiction of the Court; the comparable figure in 1987 was only 49.0 per cent, and this difference is statistically significant. At least among white Americans, Bush v. Gore seems not to have undermined the Court s legitimacy. Unexpectedly, the same is true of African Americans: There is no indication of an overall diminution of loyalty toward the Court between 1987 and There is certainly some evidence of increased opposition to the Court, but it is also clear that support for the institution grew as well, with nearly three of four blacks in 2001, compared to two-thirds in 1987, declaring that the Court should not be eliminated even if it made a series of unpopular decisions. The difference is not large, but is statistically significant, and is also reflected in the difference of means between the two surveys. On the other hand, while the percentage of African Americans unwilling to limit the Court s jurisdiction increased significantly in 2001, approaching a majority, we observe an even larger increase in the percentage favouring such restrictions on the Court. The rather substantial decline between the two surveys in the percentage of blacks who are uncertain renders comparison slightly problematic, but the difference in means between the two surveys is small. Still, African Americans seem at least moderately committed to the Supreme Court. Most significantly, we find little evidence that the Court s involvement in the election changed basic attachments to the institution. Substantial racial differences in opinions toward the Court existed in 1987 and persisted -9-

12 in Black Americans are considerably less likely than whites to express loyalty toward the Court, with, for instance, a black-white difference of 14.2 percentage points on the jurisdiction proposition in 2001 and a 10.1 percentage point difference on the institutional loyalty statement. No clear pattern of change in the differential in support between blacks and whites is apparent. Though black opinion may have polarized somewhat as a result of the presidential election, whites seem to have remained relatively loyal to the Court. We can also compare the 2001 data to a similar survey conducted in Table 3 reports these results. The 2001 figures we report in Table 3 are identical to those reported in Table 1, above. [PLACE TABLE 3 ABOUT HERE] Consider the mean scores first. In every instance, the mean is higher (more support) in 2001 than it was in The differences are not enormous, but they are in every instance statistically significant. From these data it is impossible to conclude that loyalty toward the Supreme Court plummeted after the presidential election of Indeed, the percentage of Americans saying the Court can be trusted is 12.5 percentage points higher in 2001 than it was in Americans were somewhat more likely to have an opinion toward the Supreme Court in 2001, and their opinions were more likely to be positive than in Summary This analysis yields several conclusions. First, the US Supreme Court enjoys at least a moderate degree of loyalty from the American people its reservoir of good will is certainly not shallow. Second, although blacks continue to be less likely than whites to support the Court, loyalty toward the institution in the African American community is still fairly widespread. Third, support for the Court does not seem to have been depressed by the justices involvement in the presidential election: It would be difficult indeed to conclude from these cross-sectional -10-

13 data that the basic legitimacy of the Court was threatened by the justices involvement in the 2000 election imbroglio. VIEWS OF THE COURT S OPINION IN BUSH V. GORE Americans appear to have been bitterly divided during and immediately after the 2000 election. What is less clear is how they felt about the Supreme Court and whether they regarded the Court s involvement in the election as legitimate and proper. In our survey in early 2000, we explored this issue with two questions about how our respondents felt about the Court s decision. We first asked: As you know, last month, a majority of the justices on the US Supreme Court decided that the State of Florida could not have a hand recount of the votes in the Presidential election. As a result, Al Gore conceded the election, and George W. Bush will become President of the United States. Do you think the US Supreme Court's decision in this case was fair or unfair? We followed this question with a query about the strength of the respondent s views on this matter. Three things stand out clearly from the responses to this question. First, a majority of Americans (56.2 per cent) judge the Court to have acted fairly in the dispute, even though a very large minority (41.9 per cent) thought the decision unfair. Second, nearly all of our respondents (97.2 per cent) have an opinion on this issue. Finally, opinions are firmly held, with the proportion of those holding strong views on the matter (80.5 per cent) greatly outnumbering those with more moderately held opinions (16.6 per cent). Those who disagreed with the decision are more likely to feel strongly about it than those who supported the outcome. The responses to this question confirm the sharp divisions that so characterized virtually every aspect of the

14 presidential election. Since there were so many allegations that the Court s decision was ideological, partisan, or otherwise unprincipled, we also asked the respondents their views of the criteria the Supreme Court used in making its decision. Do you think the justices of the US Supreme Court who voted to end the recount in Florida did so mostly based on the legal merits of the case or mostly based on their own desire to have Bush as the next president? We assume that most Americans expect the justices to make decisions on the basis of the legal merits of a case and disapprove of the failure to do so; indeed, we suspect that most see a decision based on the justices presidential preferences as procedurally unfair and improper. 31 Most Americans (62.4 per cent) believe that the Court based its decision on the legal merits of the case, not on the justices desire to see Bush become president. The divisions here are less stark than on the first question, with only 29.0 per cent denying that the decision was based on the case s legal merits. Not surprisingly, this question proved to be somewhat more difficult than the first, with 8.6 per cent of the respondents unable to judge the basis of the Court s decision. A fairly obvious hypothesis is that the responses to these two questions are closely connected, and, as Table 4 reveals, in fact they are. Those who believe the decision was based on the political preferences of the justices strongly tend to oppose the decision (86.4 per cent); those who assert that the decision was based on the case s legal merits are equally strongly likely to support the decision (80.0 per cent). This is a substantial relationship indeed. Although we draw no conclusion about the direction of the causality connecting these two variables, how people perceive the decision to have been made has much to do with their evaluations of Bush v. Gore. [PLACE TABLE 4 ABOUT HERE] -12-

15 The causal structure connecting institutional loyalty with judgments of the decision may be that loyalty toward the Court causes one to perceive its decisions as legal and fair loyalty acts as a frame within which to judge the opinion or perhaps perceptions that the Court acted fairly and legally in Bush v. Gore lead to enhanced loyalty toward the Court. This is the eternal dilemma of cross-sectional studies causality is inherently ambiguous. Yet from the vantage of Legitimacy Theory, untangling the causality among these variables is of the utmost importance. Investigating Causality To examine reciprocal causality between perceptions of Bush v. Gore and loyalty toward the Supreme Court, we must first develop a model of the origins of these attitudes. Earlier research on attitudes toward the Court provides some useful hypotheses. One of the strongest predictors of loyalty identified to date is support for the institutions and processes of democracy. Those who are more supportive of democracy in general are more likely to support the Supreme Court. 32 Consequently, we measured support for democratic institutions and processes along two major dimensions: (1) support for the rule of law; and (2) support for a multi-party system. We constructed indices representing each of these concepts, after confirming their unidimensionality through factor analysis. (See Appendix B for the text of all of the items used in this analysis). We hypothesize that those more loyal to the Supreme Court hold values more supportive of democratic institutions and processes. Earlier research has also strongly suggested that loyalty toward the Court is a function of exposure to its activities, as represented by awareness of and knowledge about the institution. The research literature has firmly established that to know more about the Supreme Court is to think more highly of it. 33 We employ two indicators here: (1) self-proclaimed awareness of the Supreme Court; and (2) knowledge of the Court, based on a three-item information test. We incorporate in this model several measures of satisfaction with the performance of -13-

16 the Court, especially in conjunction with the election. Bush v. Gore essentially decided the outcome of the 2000 presidential election, giving the presidency to George W. Bush. Whether one viewed this as the proper outcome may depend in part on how one views the outcome of the balloting. Thus, we included two measures of election opinions the respondent s judgment of who won the national election and of who won the election in Florida. We expect these variables to be strong predictors of the perceived fairness of the Bush v. Gore decision. Conventional wisdom indicates that partisanship structures views of the Court; 34 and, given the circumstances, this may be especially true in the aftermath of Bush v. Gore. Thus, we hypothesize that loyalty is related to party identification and to affect toward George W. Bush (as measured through a feeling thermometer). We also expect that these variables will predict reactions to the Court decision. Finally, given our findings above, we hypothesize that black Americans will extend less loyalty to the Court and will disapprove more strongly of the Court s election decision. Table 5 reports the results of regressing judgments of the fairness of Bush v. Gore and loyalty toward the Court on these independent variables. [PLACE TABLE 5 ABOUT HERE] Perhaps the most interesting finding in this table concerns the role of values in predicting both judgments of Bush v. Gore and institutional loyalty. Those who support democratic institutions and processes are significantly more likely to express loyalty toward the Supreme Court, but they are no more or less likely to perceive the decision in Bush v. Gore as fair. This is particularly noteworthy, because one of the democratic values included in the equation is support for the rule of law. Those more strongly attached to the rule of law are not necessarily more critical of Bush v. Gore. These findings confirm an important conclusion from earlier research: Institutional loyalty depends upon the relatively stable, core political values citizens hold. -14-

17 We also note the weak independent influence of partisanship on institutional loyalty: No direct relationship between party identification and loyalty exists. On the other hand, party identification is a strong predictor of the perceived fairness of the decision in Bush v. Gore. It appears that judgments of the fairness of the Court s decision and attitudes toward the Court itself are cut from different cloth. Views of the outcome of the election influence judgments of decisional fairness, but have little impact on institutional loyalty. This is important evidence of the degree to which institutional support is insulated from contemporary events. Moreover, what little effects the two electoral variables have on institutional legitimacy go in opposite directions. As we expected, those who believe Bush won nationally and in Florida are more likely to judge Bush v. Gore fair. But those judging Bush to have won nationally are (slightly) less likely to express support for the Court, an unexpected relationship that we discount since it is so weak and only achieves marginal statistical significance. The independent impact of affect toward George W. Bush is significant in both equations. It is difficult to know exactly how to understand this variable, however, in that its influence is independent of judgments of the outcome of the election and of party attachments. It is not surprising that this variable would influence judgments of the decision; that Bush supporters tend to express stronger loyalty toward the Court, irrespective of the election, most likely reflects satisfaction with the long history of conservative decisions by the Rehnquist Court. Perhaps affect for Bush captures a variety of types of variance. 35 It is also noteworthy that perceptions of who won in Florida have a considerably stronger influence on judgments of the fairness of the decision than do perceptions of who won nationally. This differential effect no doubt reflects the direct connection of the Court decision to the Florida election, as well as the confusion about who won nationally because of the biassing -15-

18 effects of the Electoral College. We also confirm the conventional wisdom concerning knowledge and information: Those who are more aware of the Court and know more about it tend to express more loyalty toward it. Both knowledge and awareness have independent influences on loyalty, with knowledge being a more powerful predictor than self-proclaimed awareness. Thus, the influence of awareness also extends beyond the effect of simple factual information, and may include, for example, attentiveness to the rituals associated with Supreme Court decision making. Finally, we note that the independent effect of race on both attitudes is statistically significant, but small. This finding is not surprising since race is strongly related to the other independent variables included in the equation (e.g., party identification, perceptions of who won the election). Attitudes toward the rule of law have a direct effect on loyalty toward the Court. But support for the rule of law may also act as a conditional variable. That is, those more supportive of the rule of law may perceive everything about the election dispute in legalistic terms, thus moderating the relationships between the independent and dependent variables. This hypothesis can be investigated with the data at hand. The Conditional Effect of Support for the Rule of Law To test the intuition that attitudes toward the rule of law have a conditional effect on the relationships depicted in Table 5, we added terms representing the interactions of attitudes toward the rule of law and each of the independent variables. In the case of loyalty toward the Supreme Court as the dependent variable, the addition of the interaction terms to the linear equation results in a statistically significant increase in explained variance. However, only one of the interaction terms achieves statistical significance: As knowledge of the Court increases, the connection between support for the rule of law and support for the Court becomes stronger. The -16-

19 slope for the rule of law variable varies from -.16 among the weakest supporters of the rule of law to.15 among those most firmly committed to the rule of law. This conditional relationship makes sense to the degree that knowledge of the Court makes it easier to see the connection between what the Court does and the cherished value of legal universalism. For no other independent variable is there a statistically significant conditional effect. In terms of the perceived fairness of Bush v. Gore, the change in explained variance brought about by the addition of the interactive terms barely achieves statistical significance (p =.04). The regression coefficients for two of the interactive variables the rule of law interactions with awareness of the Court and with perceptions of who won the national election are both significant at p =.02. As awareness of the Court increases, the relationship between support for the rule of law and perceived fairness becomes more negative. Among the most aware, there is a strong negative relationship between rule of law attitudes and judgments of the decision. Similarly, some interaction exists between rule of law attitudes and perceptions of who won the election. Among Gore supporters, the coefficient is.18; among those who are uncertain who won, it is.06; and among Bush supporters, the slope is These figures support two significant conclusions. First, rule of law attitudes are more important for forming judgments of the Court decision among Gore supporters, as one might expect. Second, among Gore supporters, greater support for the rule of law is associated with judgments that the Court was fair. None of these coefficients is particularly strong, but they nonetheless demonstrate the multiple pathways by which rule of law attitudes influenced perceptions of the Court s involvement in the dispute. Two-Stage Least Squares In order to estimate the direction of causality between perceptions of the Court s decision and institutional loyalty, we conducted a two-stage least squares analysis. The analysis is based on the hypotheses we put forth above, and uses the set of instruments drawn from the predictors of -17-

20 the perceived fairness of the opinion and loyalty to the Court (as reported in Table 5, above). The results are reported in Table 6. As hypothesized, the significant exogenous variables for perceptions of the opinion (party identification, perceptions of the national and Florida election outcomes, and affect for Bush) account for nearly all of the total explained variance (R 2 =.58); the significant hypothesized variables for loyalty (support for the rule of law, support for a multiparty system, affect for Bush, and awareness and knowledge of the Court) are also strong predictors of that variable (R 2 =.23). With only a single predictor significantly related to both of the endogenous variables, the first-stage analysis is strong enough statistically to support the second stage of the least squares. 36 [PLACE TABLE 6 ABOUT HERE] In the two-stage least squares, the findings for the variables predicting institutional loyalty and assessments of the decision differ insignificantly from those of the ordinary least squares analysis, so little further comment is required. The coefficients of greatest interest are those linking perceptions of the opinion and institutional loyalty. From the second-stage equation, these coefficients are: Opinion Fairness = Institutional Loyalty + [the exogenous variables] The coefficient for loyalty is highly statistically significant (p <.000, standard error =.13). For institutional loyalty, the equation is: Institutional Loyalty = Opinion Fairness + [the exogenous variables] The coefficient for fairness does not achieve statistical significance (p >.05, standard error =.03). These results indicate that institutional loyalty influences judgments of the fairness of the decision in Bush v. Gore, but not vice versa. For instance, according to the two-stage least squares equation, the expected opinion fairness score for those with the lowest loyalty toward the -18-

21 Court (a score of 1) is.75, while those most strongly committed to the Supreme Court have an expected value of Loyalty strongly structures perceptions of the fairness of the decision in Bush v. Gore. On the other hand, when perceived fairness is at its lowest value, loyalty is expected to be 1.56; and when fairness is at the highest value, loyalty only increases to Thus, the effect of assessments of the Court s election decision on enduring loyalty toward the institution itself is marginal indeed, reinforcing the conclusion that the Court may well have diminished its legitimacy by its ruling in Bush v. Gore, but only by a trivial amount. The reservoir of good will enjoyed by the Court greatly attenuated any negative effects of the ruling in Bush v. Gore. Indeed, institutional loyalty seems to have framed the decision, making it more palatable to those who objected to the outcome. These findings comport well with our understanding of how people update their views toward institutions. We contend that established views toward an institution influence the way in which controversial decisions are perceived and evaluated. Some think of these established views as a sort of running tally, an historical summary of how people have reacted to institutional decisions in the past. Any given adverse decision most likely does not dramatically alter these established views; but, as the statistical analysis implies, nor are reactions to highly controversial individual decisions of no consequence for how one feels about an institution. We are again attracted to the notion of loyalty. Loyalty is not obdurate; it changes, even if slowly, in reaction to people s experiences. Loyalty shapes perceptions of individual events, but individual events also shape loyalty. Our most important conclusion from this analysis is that even an enormously controversial decision like Bush v. Gore has little if any influence on institutional loyalty. -19-

22 A Theoretical Explanation of the Findings: Positivity Frames These results may reflect the bias of positivity frames when it comes to the Court (and perhaps judicial institutions in general), in the sense that exposure to courts including exposure associated with controversial circumstances enhances rather than detracts from judicial legitimacy, even among those who are disgusted with the Court s ruling. When courts become salient, people become exposed to the symbolic trappings of judicial power the marble temple, the high bench, the purple curtain, the black robes. 38 When the news media covered the Court s deliberations surrounding the election, it generally did so with the greatest deference and respect. The contrast in images of the partisan bickering in Florida and the solemn judicial process in Washington could not be more stark. No matter how one judges the outcome in Bush v. Gore, exposure to the legitimizing symbols of law and courts is perhaps the dominant process at play. Thus, the effect of displeasure with a particular court decision may be muted by contact with these legitimizing symbols. To know courts is indeed to love them, in the sense that to know about courts is to be exposed to these legitimizing symbols. One way to investigate this conjecture is to compare the views of partisans in 1987 and Following the methodology reported in Kritzer, 40 we display in Table 7 the relationship between party attachment and a two-item loyalty index (based on Table 2, above). The correlation between loyalty and party identification in 1987 is.05; in 2001, it is.20. Thus, our findings comport with Kritzer s in the sense that attitudes toward the Court seem to be more partisan following the election of 2000 but not greatly so. [PLACE TABLE 7 ABOUT HERE] The more interesting evidence has to do with whether our conclusion that legitimacy changed little between 1987 and 2001 is due to off-setting tendencies among Democrats and Republicans. That is, if Democrats reduced their support for the Court as a result of Bush v. Gore -20-

23 and Republicans increased their support, then the overall level of support would appear not to have changed. Support would, however, be more closely related to partisanship. Most important, this would be evidence that people were adjusting their loyalty on the basis of the Court s ruling on the election. Table 7 reports data of considerable relevance to the hypothesis that loyalty frames reactions to individual decisions and that unwelcome decisions contribute little to the diminution of institutional loyalty The evidence in Table 7 indicates that, between 1987 and 2001, support for the Court did not decline among Democrats, even as it increased somewhat among Independents and Republicans. In 2001, 46.6 per cent of the Democrats gave two supportive replies to our questions; in 1987, this figure was insignificantly lower (43.0 per cent). That Democrats in the aggregate seem not to have been affected by the adverse ruling in Bush v. Gore is a finding compatible with the general view that institutional loyalty inoculates against an unwelcome policy decision. The data also reveal that Republican support for the Court was boosted by the decision. Thus, these data are compatible with the conclusion that the Court profits from a bias of positivity frames in the sense that the Court gets credit when it pleases people, but that it is not penalized when its actions are displeasing. DISCUSSION AND CONCLUDING COMMENTS Several significant conclusions have emerged from this analysis. In terms of substantive politics, we have shown that the Supreme Court decision in Bush v. Gore did not have a debilitating impact on the legitimacy of the US Supreme Court. Perhaps because the Court enjoyed such a deep reservoir of good will, most Americans were predisposed to view the Court s involvement as appropriate, and therefore dissatisfaction with the outcome did not poison attitudes toward the institution. This finding is an important corrective to popular and scholarly views of the politics -21-

24 of the election. Nevertheless, no one can doubt that loyalty toward an institution is influenced by the policy outputs of that institution, at least in the long term. 41 Neither should any one doubt that loyalty toward an institution can cushion the shock of a highly controversial decision. Within any given cross-section, the causal relationship between perceptions of an opinion and loyalty is surely reciprocal. Nonetheless, Bush v. Gore seems to have had a much smaller effect on the attitudes of Americans toward their Supreme Court than many expected. The various analyses presented in this paper support the view that the weak effect of the Supreme Court s participation in the election is most likely due to pre-existing attitudes toward the Court that blunted the impact of disapproval of the Court s involvement in the election. Thus, in general, the conclusions in which we have the greatest confidence are: (1) the ruling in Bush v. Gore did not greatly undermine the legitimacy of the Court, (2) probably because the effect of pre-existing legitimacy on evaluations of the decision was stronger than the effect of evaluations on institutional loyalty, and (3) institutional loyalty predisposed most Americans to view the decision as based on law and therefore legitimate. From a more theoretical viewpoint, we have posited the existence of a bias of positivity frames when it comes to popular perceptions of courts. In most areas of political and social life, negativity predominates. We have suggested that positive reactions result from exposure to the highly effective legitimizing symbols in which courts, and the Supreme Court in particular, typically drape themselves. Further research should more rigorously investigate exactly how and under what conditions symbols are effective at legitimizing judicial institutions. Our analysis has reinforced a panoply of unanswered questions; three bear particular emphasis. First, only with more valid measures of institutional legitimacy can we make progress -22-

25 42 in unravelling the causal linkages between performance and legitimacy. We despair in particular at the use of the standard confidence measures ( confidence in the leaders of the Supreme Court ) for a variety of reasons, not the least of which is that confidence is too much affected by short-term forces. Second, we believe that cross-sectional analysis is inadequate to the task of addressing the question of how people update their attitudes toward institutions. We have pushed our cross-sectional data about as far as we dare; but answering the question of how people update their views of institutions requires a longitudinal research design in particular, a panel. Alas, such designs are much too uncommon in research on judicial politics. Finally, how do institutions build the sort of legitimacy enjoyed by the US Supreme Court? Legitimacy is extraordinarily valuable to institutions, as Bush v. Gore has so clearly shown. Understanding how institutions acquire and spend legitimacy remains one of the most important unanswered questions for those interested in the power and influence of judicial institutions. -23-

26 APPENDIX A: SURVEY DESIGN This research is based on a survey conducted in early 2001 by the Center for Survey Research (CSR) at Ohio State University. The survey is based on a typical Random-Digit-Dial (RDD) sample of the American mass public and an over-sample of African Americans. The fieldwork in the primary sample was conducted from 5 January through 19 January 2001, with 1,006 interviews completed during this period. Telephone interviewing was employed, utilizing a RDD sample purchased from Genesys Sampling Systems. The sample is representative of Englishspeaking households in the forty-eight contiguous US states (and Washington, D.C.). Within households, respondents were selected by the last birthday technique. 43 The median length of interview was about twenty minutes. Using the AAPOR standards, 44 several response rates were calculated. According to AAPOR Response Rate 5, our survey had a rate of 35 per cent; according to Response Rate 1, the rate was 26 per cent. Using AAPOR s Cooperation Rate 3, our cooperation rate was 49 per cent; modifying this rate by taking into account all households in which it is certain that an interviewer spoke with the selected respondent, the cooperation rate climbs to 78 per cent. We also surveyed an over-sample of African Americans. We sampled from census tracts in which the concentration of African-American households was 25 per cent or greater. The field work was conducted from 22 January 2001, through 12 February In all respects, the methods employed in the over-sample were identical to those employed in the primary sample. A total of 409 interviews with African-American respondents were completed in the over-sample. The response rates for the over-sample are: AAPOR Response Rate 1: 30 per cent; AAPOR Response Rate 5: 40 per cent; AAPOR Cooperation Rate 3: 55 per cent; and Modified AAPOR Cooperation Rate 3: 80 per cent. We have weighted these data to adjust for the unequal probabilities of selection (i.e., the -24-